Quantitative deterioration assessment of road bridge decks based on site inspected cracks under stagnant water

Abstract

Stagnant water on reinforced concrete (RC) decks shortens their fatigue life significantly compared to dry conditions. By using a multi-scale simulation together with the pseudo-cracking method, the remaining fatigue life of real RC bridge decks covered by stagnant water is estimated based upon their site-inspected surface crack's patterns. For quick diagnosis for deterioration magnitude of RC decks, two assessment methods are proposed. A predictive correlation of the remaining fatigue life and a mechanics-based parameter (cracks density) considering both cracks length and width is introduced as a speedy judgment for the deterioration-magnitude. For comprehensive judgment for the deterioration-magnitude, an artificial neural network (ANN) model is further introduced by means of machine learning. Bayesian regularization technique was conducted to the training scheme to reduce the misguided ANN's evaluation caused by overlearning. Finally, deck's bottom surface map of reference is introduced to show the location of comparatively problematic cracks based upon the weights assigned with the synapses of the neuron of the built ANN model.